GRA 2272 Language Processing as Organizational Cognition

GRA 2272 Language Processing as Organizational Cognition

Course code: 
GRA 2272
Department: 
Leadership and Organizational Behaviour
Credits: 
6
Course coordinator: 
Jan Ketil Arnulf
Course name in Norwegian: 
Language Processing as Organizational Cognition
Product category: 
Master
Portfolio: 
MSc in Leadership and Organisational Psychology
Semester: 
2023 Autumn
Active status: 
Active
Level of study: 
Master
Teaching language: 
English
Course type: 
One semester
Introduction

Language is arguably one of the most fascinating features of the human brain and a key topic in cognitive psychology. The cognitive processes making up the foundations of human language processing are key to the uniquely human capability for labor division, task distribution and sensemaking in organizations. Leadership, work motivation, planning and organizational learning are all dependent on language processing. Increasingly, computerized text algorithms are available to sample, analyze and map language usage in organizations. The interface between neuropsychological processes and natural language processing is already becoming the next frontier in organizational psychology. In earlier years, statistical modelling for psychological theory building and managerial applications have needed to rely on rating scales and other mapping techniques that come in addition to the linguistic interaction in organization. The topic of this course is to introduce students to modern text modeling approaches that render more direct access to cognitive and emotional psychological information. As an aspect of this, the course also introduces, explains and applies text algorithmic tools like intelligent chatbots and other software that mimics human cognition. It is important that organizational psychologists understand how uniquely human language and cognition is related to the growing use of machine-based technologies in the workplace.

This course takes the students directly to the questions of how the core topics of organizational behavior such as leadership, motivation, personality, learning and organized sensemaking can be understood and in conjunction with neurobiology and natural language processing. The course brings together foundational theory in these fields with easily accessible tools such as the chatbot ChatGPT, websites for text analysis and packages for data analysis in programming languages such as R and Python. The course focuses on how data used in organizational decision making can now be obtained through the direct processing of existing text materials instead of going methodological detours like collecting survey data with Likert scales.

The rapid advancement of these techniques has raised a number of issues from the role of human-machine interaction in advanced work, through the role of text algorithms in learning and education to questions about scientific philosophy in generating theories with predictive value in social science. The applications span increasingly wide domains in management and need to be addressed from an organizational psychological perspective, because they change work processes in the interface between humans and machines.

Examples of such organizational domains are:

  1. Automated mapping of organizational constructs such as motivation, leadership, engagement etc. that so far have only been available through survey research data.
  2. Personality profiling of applicants and others through free text.
  3. Understanding sensemaking in organizations through topic identification and extraction from texts.
  4. Understanding how human cognitive psychology can be enhanced and challenged by cognitive automatization such as Chatbot interfaces, for example between students, professors, customers, employees, and managers.
  5. Sentiment analyses assessing the attitudes of customers, employees and other stakeholders towards products, organizations and departments.
  6. Sustainability issues involved, such as decision making, diversity and technology acceptance and literacy in the workplace.
  7. Ethical issues involved in digital constructions of social realities, such as biases, privacy and power distribution.

This development of low-cost cognitive tools is changing the way we understand language as maps of organizations and emphasizes the need to understand the psychology of language. The three most important theoretical underpinnings are: 1) Modeling of human cognition in the interface between neuropsychology and the digital technologies for natural language processing (e.g., Landauer), 2) Sensemaking in organizations as a way of understanding continuous organizing as an ongoing language game (e.g., Weick), and 3) Social constructionism as a foundation for how our social realities are constituted and upheld by linguistic practices (e.g., Berger & Luckmann).

The aim of this course is to unify these three strains of organizational psychology with state-of-the-art computational tools in text analysis and management. Given the rapid advancement of technologies, the course is also building on contributions from complex adaptive systems theory. The course addresses specifically the psychological interface between human cognitive psychology, computational tools and organizational theory and practice. It does not cover organizational communication or strategic aspects of digitization, nor does it give the students any detailed instruction in writing computer code.

Learning outcomes - Knowledge

By the end of the course the candidate:

  • has advanced knowledge about the role of language in human cognition, including the neurobiological and socio-cultural determinants of language.
  • understands why and how computational tools can be used to analyze and simulate language and texts, model core psychological constructs such as leadership, motivation and personality, including how digital systems may emulate, assist or enhance human psycholinguistic performance.
  • has advanced knowledge about state of the art applications of text analytical instruments in organizations
  • has specialized knowledge about existing and possible applications of text analysis in managerial decision making
  • Understands how language usage is affected by diversity in groups of language users, such as between cultures, demographic subgroups, professional groups or between nations
Learning outcomes - Skills

By the end of the course the candidate:

  • Is able to make reflected decisions on how and when to use automated text algorithms to support psychological work processes, including knowledge generation in higher education.
  • is able to use a number of open-source digital tools to collect, browse and analyze texts for decision making in organizations
  • can make realistic assessments of text in organizations as sources of information
  • is able to search and assess rapidly developing new tools, extensions and packages that improve and expand methods of text analysis
  • is able to take part in discussions about automated text procedures in management decision making, organizing work processes and assessment of HR topics
  • is realistic about the limitations of text processing
  • is able to explain and visualize data from text analysis
General Competence

By the end of the course the candidate:

  • is able to address how human language processing is contributing to the mapping, construction and operation of organizations.
  • Is able to critically examine and use different theoretical approaches to text analysis, with reference to neurobiological, socio-cultural and information theoretical explanations.
  • can critically reflect upon, communicate, and discuss a selection of theories and research findings from digital text analysis and applications in organizational settings.
  • is able to span the bridge between computer scientists, professionals in organizations and practicing managers
  • can apply knowledge and skills to contribute to novel thinking and innovation processes crucial to designing human-machine interfaces in organizations.
Course content

This course starts out with a direct introduction to some up-to-date applications of algorithmic text analysis, such as ChatGPT and other open source tools and research findings. The course then takes the students along a path to understand how and why these technologies are possible, keeping a focus on their possible uses and limitations. The course is a continuous mixture of lectures, discussions and student exercises.

The progress of the course contents is structured as follows:

1) Organizations as talking machines: ChatGPT as a window to organized work, and understanding language as a map of shared human realities.

2) Psycholinguistcs: What is actually human, natural language and how does the brain interact with culture to produce it?

3) Information theory: Introduction to the physical rules that unify brains, language and computers.

4) The genuinely human side to language: Pragmatics of human communication, reasoning theory and why natural language works better for lawyers than for scientists, including biases and power.

5) The role of language for change and stability in organizations: social construction, sensemaking, change poetry and the multi-lingual work environment with implications for diversity.

6) Digital language processing: Topics, wordclouds, sentiments and semantic analyses. Easy access to text analysis and the possible advanced options.

7) Statistical semantics: Modelling constructs such as leadership, motivation and personality. Predicting human behavior using available algorithmic tools. Constructs as intersection of social practice and information theory.

8) A touring of contemporary text analysis in organizations: Updated applications from the corporate landscape.

Teaching and learning activities

A key premise to this course is that digital text tools are in continuous development as the course itself is being planned. The basic pedagogical idea of this course is therefore to introduce freely available tools to the students at the beginning, and to encourage the students to use these tools as a way to prepare and engage throughout the entire course. Also, the course plans to arrange 1/3 of the course hours as a-synchronous learning activities, where the students will work individually and in groups to try out and apply the tools used in the course.

However, the main ambition for the course is that the students will be able to understand the gap between machine generated text analysis and human cognition. To achieve this, there will be interactive lectures and discussions to stimulate the students to detect and explain how machine implemented cognition is similar to, but also different from human cognition. The course will require students to be able to voice informed theoretical, technical and epistemic concerns about the use of digital text tools without the immediate input of such tools. Genuine human face-to-face interaction and the ability to read and understand relevant literature is therefore just as much a part of this course as the digital technologies themselves.

This topic is so new that there is no single textbook that covers the topics. The course only foresees two compulsory books: One recent on psycholinguistics and a classic introduction on information theory. The rest of the topics are covered through updated research or review articles and documentation of programming tools and packages.

The digital tools to be used in this course are listed below but not limited to these as new possibilities emerge all the time. The key to technologies of choice is that this course does not require the students to write complicated programming code. The course is open to students at various levels of programming skills, such that programming enthusiasts should have the same options to develop their perspectives as students without any previous knowledge in this field. The main technologies that will be used are:

ChatGPT (or similar). This is an open-source multi-purpose chatbot tool that may e.g. be used for explanations, personalizing the learning journey and also for help in writing code.

Voyant: This is an open-source web-based tool with multiple functionalities in analyzing and illustrating features of texts.

Google nGram: A powerful tool to analyze the historical buildup of many major languages

http://wordvec.colorado.edu/: An easy-to use open-access engine to compute semantic patterns in text data

https://semanticexcel.com/: A slightly more complicated but open-access tool to compute semantic patterns in data, much used to predict personality and clinical phenomena and published in high-quality journals.

https://rosenbusch.shinyapps.io/semantic_net/: An open access, easy to understand website to compare psychological measurement instruments and scales.

https://wordvis.com/: A visualization tool for the lexical WordNet database at Princeton

Packages in R and Python such as

TidyText: Great for sentiment analysis

LSAfun: Create your own expert system to answer questions based on the documents that you feed it.

Software tools
Software defined under the section "Teaching and learning activities".
Additional information

Please note that while attendance is not compulsory in all courses, it is the student’s own responsibility to obtain any information provided in class.

Qualifications

All courses in the Masters programme will assume that students have fulfilled the admission requirements for the programme. In addition, courses in second, third and/or fourth semester can have specific prerequisites and will assume that students have followed normal study progression. For double degree and exchange students, please note that equivalent courses are accepted.

Disclaimer

Deviations in teaching and exams may occur if external conditions or unforeseen events call for this.

Required prerequisite knowledge

This course aims at students on a Mastere of Science level who are interested in the intersection of language, human cognition, organizations and digital text algorithms. Other than a previous bachelor's degree in relevant topics, no specific previous knowledge is required as long as the student qualifies for a master's level course. Background interest in natural language processing or cognition is a benefit, but not a necessity.

Exam categoryWeightInvigilationDurationGroupingComment exam
Exam category:
Submission
Form of assessment:
Written submission
Exam code:
GRA 22721
Grading scale:
ECTS
Grading rules:
Internal examiner
Resit:
Examination when next scheduled course
100No2 Month(s)Individual Individual written assignment.
Exams:
Exam category:Submission
Form of assessment:Written submission
Weight:100
Invigilation:No
Grouping (size):Individual
Duration:2 Month(s)
Comment:Individual written assignment.
Exam code:GRA 22721
Grading scale:ECTS
Resit:Examination when next scheduled course
Type of Assessment: 
Ordinary examination
Total weight: 
100
Student workload
ActivityDurationComment
Teaching
24 Hour(s)
24 hours of synchronous teaching.
Teaching
12 Hour(s)
12 hours of asynchronous teaching.
Prepare for teaching
36 Hour(s)
Group work / Assignments
28 Hour(s)
Student's own work with learning resources
60 Hour(s)
Students own work with preparations for exams.
Sum workload: 
160

A course of 1 ECTS credit corresponds to a workload of 26-30 hours. Therefore a course of 6 ECTS credits corresponds to a workload of at least 160 hours.